Performance of Deep Learning Pickers in Routine Network Processing Applications
نویسندگان
چکیده
Abstract Picking arrival times of P and S phases is a fundamental time-consuming task for the routine processing seismic data acquired by permanent temporary networks. A large number automatic pickers have been developed, but to perform well they often require tuning multiple parameters adapt them each dataset. Despite great advance in techniques, some problems remain, such as difficulty accurately pick waves earthquake recordings with low signal-to-noise ratio. Recently, phase based on deep learning (DL) shown potential event identification arrival-time picking. However, general adoption these methods monitoring networks has held back factors availability well-documented software, computational resources, gap knowledge methods. In this study, we evaluate recent available DL data, comparing performance several neural network architectures. We test selected using three datasets different characteristics. found that analyzed (generalized detection, PhaseNet, EQTransformer) tested cases. They are very efficient at ignoring large-amplitude transient noise picking waves, difficult even experienced analysts. Nevertheless, varies widely terms sensitivity false discovery rate, missing significant percentage true picks others producing positives. There also variations run time between pickers, requiring resources process datasets. spite drawbacks, show can be used efficiently obtain results comparable or better than current standard procedures.
منابع مشابه
analysis of power in the network society
اندیشمندان و صاحب نظران علوم اجتماعی بر این باورند که مرحله تازه ای در تاریخ جوامع بشری اغاز شده است. ویژگیهای این جامعه نو را می توان پدیده هایی از جمله اقتصاد اطلاعاتی جهانی ، هندسه متغیر شبکه ای، فرهنگ مجاز واقعی ، توسعه حیرت انگیز فناوری های دیجیتال، خدمات پیوسته و نیز فشردگی زمان و مکان برشمرد. از سوی دیگر قدرت به عنوان موضوع اصلی علم سیاست جایگاه مهمی در روابط انسانی دارد، قدرت و بازتولید...
15 صفحه اولApplications | Uncertainty in Deep Learning
In previous chapters we linked stochastic regularisation techniques (SRTs) to approximate inference in Bayesian neural networks (NNs), and studied the resulting model uncertainty for popular SRTs such as dropout. We have yet to give any real-world applications stemming from this link though, leaving it somewhat in the realms of “theoretical work”. But a theory is worth very little if we can’t u...
متن کاملDeep learning evaluation using deep linguistic processing
We discuss problems with the standard approaches to evaluation for tasks like visual question answering, and argue that artificial data can be used to address these as a complement to current practice. We demonstrate that with the help of existing ‘deep’ linguistic processing technology we are able to create challenging abstract datasets, which enable us to investigate the language understandin...
متن کاملFlexible Deep Neural Network Processing
The recent success of Deep Neural Networks (DNNs) has drastically improved the state of the art for many application domains. While achieving high accuracy performance, deploying state-of-the-art DNNs is a challenge since they typically require billions of expensive arithmetic computations. In addition, DNNs are typically deployed in ensemble to boost accuracy performance, which further exacerb...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Seismological Research Letters
سال: 2022
ISSN: ['0895-0695', '1938-2057']
DOI: https://doi.org/10.1785/0220210323